Huw,
Can you look at the debug log after applying the SupportInfo Patch in the bug#6341. I think QI should be done for IXMLDocument not IXMLNode. I will try to do it for IXMLNode as well.
Thanks, VJ
On 10/11/06, Huw Davies huw@codeweavers.com wrote:
On Wed, Oct 11, 2006 at 12:55:41AM -0400, Vijay Kiran Kamuju wrote:
I will wait on sending tests on this one. Until the first one gets in (http://www.winehq.org/pipermail/wine-patches/2006-September/030659.html)
Ok, so let's look at that one.
I think the problem here is that ISupportErrorInfo should hang off the underlying node object (in node.c) rather than domdoc itself. All of the objects that inherit from IXMLDOMNode will then get ISupportErrorInfo for free rather than having to implement it for each node type (you'll obviously need to forward the QI through to the node implementation). It would be worth checking that QI on, for example, an attribute node does indeed return ISupportErrorInfo.
Another interesting thing to look at (although you can't really write a test for this) is to look at whether the pointer you get back from QI(ISupportErrorInfo) is 'close' to the pointer you get for QI(IXMLDOMDocument) or QI(IXMLDOMNode). If I'm right, it should be close to the latter, in your implementation it'll be close to the former.
Huw.